Display technologies are emerging which have importance for a variety of applications. For example, various graphical user interfaces and displays have been developed for personal computing, financial services applications, etc. Recent advances in hardware and software technologies enable the development of powerful graphical user interfaces.
Various types of process control systems are presently in use, such as for control of processes operable under control of a single variable to processes controlled using controllers capable of controlling multiple variables. Control of a process is often implemented using microprocessor-based controllers, computers, or workstations which monitor the process by sending and receiving commands and data to hardware devices to control either a particular aspect of the process or the entire process as a whole. For example, many process control systems use instruments, control devices, and communication systems to monitor and manipulate control elements, such as valves and switches, to maintain one or more process variable values (e.g., temperature, pressure, flow, and the like) at selected target values. The process variables are selected and controlled to achieve a desired process objective, such as attaining a safe and efficient operation of machines and equipment utilized in the process. Process control systems have widespread application in the automation of industrial processes such as, for example, the processes used in chemical, petroleum, and manufacturing industries.
In recent years, advanced process control systems for controlling multivariable processes have been developed. For example, one type of process control is based on configuring or programming advanced controls based on engineer(s) knowledge (e.g., incorporating feed forward, signal selection, and calculation blocks) to continually push a process plant toward some known operating state. Another type of advanced process control is model-based predictive control. Model-based predictive control techniques have gained acceptance in the process industry due to their ability to achieve multivariable control objectives in the presence of dead time, process constraints, and modeling uncertainties.
In general, model-based predictive control techniques include algorithms which compute control moves as a solution to an optimization problem for minimizing errors subject to constraints, either user imposed or system imposed. A model-based predictive control algorithm can be generally described with reference to a multivariable process. Generally, the model-based predictive control includes two major portions: first, an optimization program is used to define the best place to run the process at steady state, and, second, a dynamic control algorithm defines how to move the process to the steady state optimum in a smooth way without violating any constraints. For example, at a specified frequency, e.g., every minute, the optimizer looks at the current state of the process and calculates a new optimum. From the optimizer, the controller knows where process variables should be in the final steady state. The control algorithm then calculates a dynamic set of changes for the process variables to move the process in a smooth way to the steady state with no dynamic violations of constraints. For example, 60-120 control moves may be calculated out into the future for a process variable. Generally, one of the calculated control moves is implemented and the rest thrown away. These steps are then reiterated. The control objective for the model-based predictive control is generally to provide for optimum controlled variables through calculation using a model based on economic values.
Model-based predictive control is performed using products available from several companies. For example, model based predictive control is performed by a Dynamic Matrix Control (DMC) product available from Aspen Tech (Cambridge, Mass.), and by a Robust Multivariable Predictive Control Technology (RMPCT) product available from Honeywell Inc. (Minneapolis, Minn.) which is a multi-input, multi-output control application product that controls and optimizes highly interactive industrial processes such as when used in suitable automated control systems.
Generally, a model-based predictive controller contains three types of variables; namely, controlled variables (CVs), manipulated variables (MVs), and disturbance variables (DVs) (sometimes also referred to as feed forward variables (FFs)). Controlled variables are those variables that the controller is trying to keep within constraints. Further, it may also be desirable to minimize or maximize some of the controlled variables (e.g., maximize the feed throughput process variable). Manipulated variables are those variables, such as valves, that the controller can open and close to try to achieve an objective of the controller (e.g., maximizing feed throughput) while maintaining all of the controlled variables within their constraints. Disturbance variables are those variables that can be measured, but not controlled. Disturbance variables assist the controller by providing needed information such as information regarding certain factors, e.g., outside air temperature. The controller can then recognize how such factors will affect other process variables in the controller, so as to better predict how the plant will react to measured disturbances.
A user of the model-based predictive controller (e.g., an engineer, an operator, etc.) has conventionally been provided with various types of information regarding the various process variables including information concerning the controlled variables, manipulated variables, and disturbance variables. For example, information such as predicted values, current values, and other relational information of variables relative to other variables has been provided to a user in the past by way of various interfaces and displays. The user can monitor such information and interact with the controller in various ways. For example, the user can turn the controller on and off, take individual process variables in and out of control, change various types of limits placed on process variables contained in the controller (e.g., change low or high limits for individual process variables), change the model of the controller, etc.
However, in order for the user to monitor the overall health of the controller effectively, and to interact with the controller in the required manner (e.g., changing limits of process variables), the user must be presented with suitable controller information. For example, an operator monitoring the controller should be presented with information regarding the relationship between manipulated variables and controlled variables, the limits to which process variables are constrained, the current values of the various process variables, etc. Such information should be presented in such a manner that a user can effectively understand the performance of the process and, for example, be able to detect and solve problems in the process. Although various types of screen displays have been used to present information regarding the controller to a user (e.g., those described in the Honeywell product publication entitled “Robust Multivariable Predictive Control Technology-RMPCT Users Guide for TPS (6/97) hereby incorporated herein by reference in its entirety and hereinafter referred to as “Honeywell Users Guide”) such that the user can monitor and manipulate parameters related to one or more process variables in the process being controlled thereby, the effectiveness of such an interface has been lacking and the users may have difficulties performing the required monitoring and control functions.
For example, one difficulty of monitoring multiple dynamic process variables in parallel is that generally a large amount of screen real estate needs to be devoted to the presentation of textual data with respect to such process variables. For example, this is particularly a problem facing operators of nuclear, chemical, and petrochemical plants where the number of dynamic process variables is large. In general, a conventional solution to this multivariable monitoring problem is the use of trend history plots that display the historical behavior of one or more variables. However, this approach is still too real estate intensive in that it requires a great deal of space to display multiple trend history plots in parallel even for just a few process variables. As such, users are typically forced to access at least some of the trend history plots for the process variables in a serial manner.
Further, for example, a user in a model-based predictive control process must be able to deduce potential causes of observed controlled variable changes and assist the users in predicting the effects of any planned manipulated variable manipulations, e.g., change of constraints or limits for a manipulated variable. One particularly beneficial screen display currently used for such analysis is a matrix table that displays a gain relationship between controlled variables and manipulated variables. For example, a gains matrix screen displaying gain values is currently available as shown in the Honeywell Users Guide. However, such displays do not provide adequate information and tools to use the matrix screen to support the user in problem solving tasks. In fact, generally, only process engineers and not operators of the controllers make frequent use of the tables.
Yet further, for example, model-based predictive controllers generally are constraint-based tools as are various other controllers, e.g., the controllers attempt to control a process within certain constraints or limits defined for process variables being controlled. The use of such constraint-based techniques for controlling the process presents the problematic task of being able to monitor or keep track of the relationships between the various constraint limits and the current values for one process variable or a multiple number of process variables. For example, in a model-based predictive controller, engineering hard limits, operator set limits, engineering physical limits, and/or various other limits may be specified for a number of different process variables. A user is generally required to monitor the relationships of a large number of process variables. Traditionally, information to carry out such monitoring is by presentation of such information in textual form. For example, a user is presented with tabular values representative of engineering high and low hard limits in addition to the current value for a process variable. The user is then required to read the text and formulate the relationship between the relevant limits and current value. When monitoring a large number of such process variables, the task of formulating such relationships is difficult.
In addition, for example, a user may be required to effectively monitor and manipulate parameters for a process variable, e.g., the setting of operator high and low limits for a process variable. Currently, interface techniques used to present information to the user and provide the user a way of changing one or more parameters of a process variable have been ineffective. For example, typically a user relies primarily on tabular presentation of data with respect to a particular process variable, e.g., color-coded tabular presentation of textual material. However, in one particular case, some graphical elements have been used to show one or more subsets of information, such as limits and current values, with supporting text, for use in monitoring and manipulating a process variable. However, such approaches suffer from at least three problems. First, they are difficult to use, either because they necessitate extensive cognitive manipulation of quantitative data or because they are incomplete in their integration. For example, when some graphics have been used with textual material, the graphics have not effectively presented such information to the user. For example, a graph including a separate pair of lines indicating limits for a process variable, a separate bar representing operator set high and low limits for the process variable, a separate line representing a present value of the process variable, and clamping limits within the other limits have been used to display characteristics of the particular process variable. However, such separate display of the elements lacks integration for easy monitoring of the process variable. Second, by having the different limit relationships displayed independently, valuable screen real estate is used up making it impossible to show more than just a few process parameters at a time. This again forces a user to do serial comparisons across several variables. Third, none of the existing graphical approaches allow for direct manipulation of the variable limits. In other words, the user must change limits indicated by using a separate screen or separate textual information.
As indicated above, the displays used to convey information to a user for monitoring and manipulation of process variables, e.g., process variables of a controller providing control of a continuous multivariable production process, are not effective. For example, one particular problem involves the use of a great deal of textual information which requires the user to formulate relationships between different process variables of the controller (e.g., formulate relationships between current values and process limits, formulate relationships from the textual matter between trends of multiple process variables, etc.). Yet further, such conventional displays which attempt to provide adequate information for a user, e.g., trend plots, textual information, etc., require an undesirable amount of screen real estate.